Building the Parsed Corpus of Scottish Correspondence

The Parsed Corpus of Scottish Correspondence (PCSC) is the first syntactically parsed corpus of Scots, annotated according to the Penn Parsed Corpora of Historical English (PPCHE) framework (Kroch, 2020). Developed from the Helsinki Corpus of Scottish Correspondence, 1543–1747 (Meurman-Solin & VARIENG, 2017), the corpus provides 270,553 words of manually corrected syntactic annotation together with rich metadata of, e.g., authors, recipients, chronology, geography, and scribal practice (e.g., whether the letter is autograph and what type of script is used). This paper discusses the methodological decisions involved in adapting an existing diplomatic corpus for syntactic annotation, including data preparation, automatic Part-of-Speech (PoS)-tagging, syntactic parsing, and manual correction. Particular attention is given to the challenges posed by non-standard and highly-variable orthography, editorial markup, as well as the absence of resources designed for Scots, and to the strategies adopted to facilitate annotation efficiency and consistency while preserving the linguistic characteristics of the source texts. Hence, beyond documenting the creation of the PCSC, the discussion highlights methodological considerations relevant to the development of parsed historical corpora more generally, particularly for under-resourced varieties. The paper outlines the annotation workflow and demonstrates how the resulting corpus supports reproducible syntactic searches using CorpusSearch2 (Randall, 2010; Randall et al., 2004). The corpus expands the range of PPCHE-type corpora available for comparative historical syntax and provides a resource for future quantitative research on the history of Scots.